Author Topic: Building a 7 decade voltage calibrator  (Read 59883 times)

0 Members and 1 Guest are viewing this topic.

Offline sync

  • Frequent Contributor
  • **
  • Posts: 799
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #50 on: October 07, 2013, 03:24:24 pm »
A KV divider will be difficult or very expensive. The resistors must be very good. I expecting the KV pot in the video costs several 1000 bucks. A DAC, PWM or AN86 should be easier and much cheaper. I experimenting with PWM at the moment with promising results. Attached a measurement from it.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #51 on: October 07, 2013, 04:05:41 pm »
Quote
I experimenting with PWM at the moment with promising results.

I'm working on a circuit too, thanks to Andreas who showed me his modified EDN version with 30bit resolution (2bit overlap) when he was at my lab for reference measurements. I also plan some modifications like the use of LTC2057 instead of LTC1151, AD586 or MAX6250 or optional LM399 with LTC1043 and LTC2057 for ~5,2V.
Like suggested by Andreas I will use the ATMEGA1284 with a 6,5536MHz crystal. The advantage is that there is no internal pll circuit for the clock.
I'm still drawing, but it will be a mixture of SMT and THT.
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Spikee

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: nl
Re: Building a 7 decade voltage calibrator
« Reply #52 on: October 07, 2013, 04:06:11 pm »
A long time ago i had the idea to build a kv divider. But it is a lot of work + expensive + time consuming.  And if you eventually finish it , who says it will perform good ?

I agree that AN86 dac method is better.
Freelance electronics design service, Small batch assembly, Firmware / WEB / APP development. In Shenzhen China
 

Offline sync

  • Frequent Contributor
  • **
  • Posts: 799
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #53 on: October 07, 2013, 04:44:20 pm »
I'm working on a circuit too, thanks to Andreas who showed me his modified EDN version with 30bit resolution (2bit overlap) when he was at my lab for reference measurements.
Interesting. Do you have any stability issues with the resistor ratio?

I using a very crude approach. A 10 bit PWM clocked at 16MHz from an Arduino and dithering the lower 14 bits. A 74HC14 as driver which is directly powered from a LT1021C 5V reference. Then a passive second-order LPF and a LTC1050 as buffer. I was surprised how well it performs.
 
The following users thanked this post: l3VGV

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #54 on: October 07, 2013, 04:47:52 pm »
Quote
I agree that AN86 dac method is better.

And I agree that the use of an AD5791 is more simple than AN86.
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Spikee

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: nl
Re: Building a 7 decade voltage calibrator
« Reply #55 on: October 07, 2013, 04:53:01 pm »
Quote
I agree that AN86 dac method is better.

And I agree that the use of an AD5791 is more simple than AN86.

And I can confirm that AD is out of samples (AD5791).
« Last Edit: October 07, 2013, 05:17:43 pm by Spikee »
Freelance electronics design service, Small batch assembly, Firmware / WEB / APP development. In Shenzhen China
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #56 on: October 07, 2013, 05:32:13 pm »
Quote
And I can confirm that AD is out of samples (AD5791).

Do you really wonder that AD don't sample a 80,-€ plus VAT part?
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Spikee

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: nl
Re: Building a 7 decade voltage calibrator
« Reply #57 on: October 07, 2013, 05:44:36 pm »
Quote
And I can confirm that AD is out of samples (AD5791).

Do you really wonder that AD don't sample a 80,-€ plus VAT part?

I have sampled more expensive parts before and it was no problem.

Backorder
Est Avail:
2013-11-08
Freelance electronics design service, Small batch assembly, Firmware / WEB / APP development. In Shenzhen China
 

Offline quantumvolt

  • Frequent Contributor
  • **
  • Posts: 395
  • Country: th
Re: Building a 7 decade voltage calibrator
« Reply #58 on: October 07, 2013, 07:45:06 pm »
Quote
(General)
Hello Gerry,
Gerry. ....
hi Gerry,
......


Strangely, the OP has not a single post here. He must be busy with his 1% resistor boards...

This thread has enough momentum to get a life of it's own. I am very interested in the different ideas presented. Nice to see that there is activity here again  ;D
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3351
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #59 on: October 07, 2013, 08:27:33 pm »
Quote
the *3/4 is then the homework.

Okay, got it. Here's the answer:

Have you simulated it?
I guess it has a not so good settling time.
And perhaps dependancy on tolerances,
because the main law of the capacitors is not met:
in the one state the capacitors are in series and in the other they are in parallel.

With best regards

Andreas
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #60 on: October 07, 2013, 08:29:26 pm »
Quote
I have sampled more expensive parts before and it was no problem.

I don't understand your problem to pay for something. It's not the task of AD or any other company to supply you with free samples, it's a friendly option. The people that work there also want to be paid.
Those all-grabbing people are the reason that many companies don't offer free samples anymore. The samples are for engineering purpose with hopefully big quantaties in series production and not for your homebrew hobby stuff. Think about that ;)
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3351
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #61 on: October 07, 2013, 08:32:34 pm »
A KV divider will be difficult or very expensive. The resistors must be very good. I expecting the KV pot in the video costs several 1000 bucks. A DAC, PWM or AN86 should be easier and much cheaper. I experimenting with PWM at the moment with promising results. Attached a measurement from it.

Around 9 ppm nonlinearity is not too bad on the first try.
My first cirquit was around 13 ppm. With carefully adjustment I could get around 3 ppm.
Which analog switches do you use for the PWM output?
Which Instrument do you use for 1ppm linearity measurement?

With best regards

Andreas
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #62 on: October 07, 2013, 08:33:55 pm »
Quote
Have you simulated it?

Yes I did.
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline sync

  • Frequent Contributor
  • **
  • Posts: 799
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #63 on: October 07, 2013, 08:56:39 pm »
Which analog switches do you use for the PWM output?
Which Instrument do you use for 1ppm linearity measurement?
I use a 74HC14 as driver which is powered from a LT1021C reference. The DMM is an old HP 34401A with unknown calibration status. It's A/D linearity is specified as 2ppm of reading and 1ppm of range.  So it's not good enough but the best i have. I took the average of 10 reading at 100 PLC.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3351
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #64 on: October 07, 2013, 09:22:40 pm »
Quote
Have you simulated it?

Yes I did.

I think its time to show my proposal:

The first half of LTC1043 (the lower) is a normal 2:1 divider from the datasheet creating 3.5V (at PIN S2A) out of 7V.
The second half is also a 2:1 divider but switched between input and the 3.5V output of the first device.
So the halved 7V-3.5V = 1.7V are added to the output of the first LTC giving 5.2V in total.
To optimize settling time the output capacitor of the first LTC  is set to 1.5uF.

BTW. there should be a 2nd solution which subtracts the 1.7V from the 7V instead of adding it to the 3.5V.
But I fear that there could be some overshoot to 7V at the output when the cirquit is switched on.

With best regards

Andreas



 
The following users thanked this post: mycroft

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #65 on: October 08, 2013, 07:01:56 am »
I agree that your circuit solution settles three times faster.
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: Building a 7 decade voltage calibrator
« Reply #66 on: October 08, 2013, 11:57:42 am »
Do you guys have an idea how they compare in terms of precision? As in how many ppm for the 3/4 circuit from Andreas compared to the one from branadic?
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3351
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #67 on: October 08, 2013, 09:22:12 pm »
Do you guys have an idea how they compare in terms of precision? As in how many ppm for the 3/4 circuit from Andreas compared to the one from branadic?

Do you mean precision (repeatability over time + temperature) or accuracy (actual deviation from theoretical value)?
I am not sure whether the values of the datasheet for a 2:1 divider of 1ppm are really accuracy or only precision.

My 2:1 dividers built according to the datasheet have a typical gain of around -9ppm against the ideal value.
But I do not know which (teflone ?) capacitors are used for the 1ppm value.

I use small WIMA MKS-02 mylar capacitors which could be surely improved by low leakage types or even polypropylene.
Further factors on accuracy are the PCB + LTC leakage + charge injection currents and the input bias current of the buffer amplifier at the output.

Interestingly the gain is higher if the pins of the LTC1043 are flipped from the right to the left side. In this case I measure around -2ppm gain which corresponds much more to the calculated error budget of 3.5ppm for the leakage currents with my components.
So I have the suspicion that the charge injection compensation circuit of the LTC1043 works in the wrong direction with the pinning of the LTC datasheet.

With best regards

Andreas

 
The following users thanked this post: cellularmitosis

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #68 on: October 08, 2013, 10:19:16 pm »
Quote
I use small WIMA MKS-02 mylar capacitors which could be surely improved by low leakage types or even polypropylene.

One of the major problems. The LTC1043 comes in a SO18 package (okay, a DIP18 version is also available) on the other hand polypropylene capacitors are huge, for example the WIMA MKP-10 comes with 27,5mm pitch. Teflon capacitors with 1µF or more are no option, as they are not available for realistic prices. NPO ceramic capacitors with 1µF and more are not available.
So what is the right choice?

I found PPS capacitors in surface mount technology:

http://de.farnell.com/panasonic/ecpu1c105ma5/kondensator-1uf-16v/dp/9694315

Any suggestions to that? PPS is know to have small leakage, small dissipation factor and therefor good hf characteristics.

"You need low leakage types, primarily capacitors with outer film that can be tied to ground."

Are there such capacitors available?
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3351
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #69 on: October 09, 2013, 06:25:23 am »
Hello branadic,

when you look at the datasheet of your tiny 1210 1uF PPS capacitor you will find a time constant of 300s as isolation resistance. That is in the range of a X5R ceramic capacitor.

I have already used WIMA 2220 PPS capacitors together with WIMA MKS-02 in a 3:1 divider cirquit.
http://www.farnell.com/datasheets/1674204.pdf
The PPS have >1250s time constant
The MKS-02 have also >1250 s
Standard mylar capacitors have >3000s at least at higher voltages.
Good polypropylene have >30000 s (and are available from 15 mm pitch)
http://www.kemet.com/kemet/web/homepage/kechome.nsf/vapubfiles/F3294_PHE426.pdf/$file/F3294_PHE426.pdf

So think BIG with precision components.

Since modern capacitors are multilayer low inductance it will be hard to find some (rectangular) with a outer film.
But you can think of copper taping the capacitor or use axilar types.

With best regards

Andreas
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2476
  • Country: de
  • Sounds like noise
Re: Building a 7 decade voltage calibrator
« Reply #70 on: October 09, 2013, 07:29:06 am »
Quote
So think BIG with precision components.

I can't, I've studied microsystemtechnology, not heavy weight equipment construction. If it can be done small do it :)
What I can fairly accept is something like caps from Vishay BC Components with pitch 5,00mm:


100nF: http://www.digikey.de/product-detail/de/BFC241641004/BC2054-ND/502834

and 15,00mm pitch:

1µF: http://www.digikey.de/product-detail/de/BFC241671005/BC2076-ND/502856
« Last Edit: October 09, 2013, 07:58:21 am by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline quantumvolt

  • Frequent Contributor
  • **
  • Posts: 395
  • Country: th
Re: Building a 7 decade voltage calibrator
« Reply #71 on: October 28, 2013, 08:14:37 pm »
IMO it is somewhat unrealistic to build a 7 decade Volt-box, but I will try to add two effective bits to the LTC2756 18 Bit Softspan DAC (+-1 LSB Error). This can be done (as in the AN86) by paralleling / attenuated cascading a second DAC or maybe by using a LTC1043 capacitive switch divider (two divisions by 2 adds 2 bits (if I am not wrong)).





http://www.linear.com/product/LTC2756
 

Offline sync

  • Frequent Contributor
  • **
  • Posts: 799
  • Country: de
Re: Building a 7 decade voltage calibrator
« Reply #72 on: October 28, 2013, 09:35:13 pm »
IMO it is somewhat unrealistic to build a 7 decade Volt-box, but I will try to add two effective bits to the LTC2756 18 Bit Softspan DAC (+-1 LSB Error). This can be done (as in the AN86) by paralleling / attenuated cascading a second DAC or maybe by using a LTC1043 capacitive switch divider
With ADC feedback like AN86?

Quote
(two divisions by 2 adds 2 bits (if I am not wrong)).
What do you divide?
 

Offline quantumvolt

  • Frequent Contributor
  • **
  • Posts: 395
  • Country: th
Re: Building a 7 decade voltage calibrator
« Reply #73 on: October 28, 2013, 10:08:35 pm »
IMO it is somewhat unrealistic to build a 7 decade Volt-box, but I will try to add two effective bits to the LTC2756 18 Bit Softspan DAC (+-1 LSB Error). This can be done (as in the AN86) by paralleling / attenuated cascading a second DAC or maybe by using a LTC1043 capacitive switch divider
With ADC feedback like AN86?

Quote
(two divisions by 2 adds 2 bits (if I am not wrong)).
What do you divide?


Yes - ADC feedback. As I now have the LTC2400 ADC running - and in addition have built the Linduino clone (expanded Arduino with a C++ library called LTSketchbook) to control / read all LT devices, all I need is another DAC (16 bit should be OK - I have one). I do not have to copy the Code Comparator PIC running ASM.

I will try several references. The LTC2756 is able to output 0-10 volt based on the 5 Volt LTC6655-5 reference.

[What will I divide?] Alternatively I can make a 2 Bit DAC from a LM399 board ouputting 10 Volt. Another possibility for reference will be 2 or 4 stacked LT6655 5V or 2.5V on separate batteries. This device will with quality relays be a programmable 2 Bit DAC with resolution 2.500.... Volt. I can then sum the output of this 2 Bit DAC to the output of the LTC2756 configured as a 0-2.500.... DAC. Since all devices are digitally controlled I can also have ADC feedback for a configuration like this.

Me thinks ... Flaw detection and ideas are welcome.
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: Building a 7 decade voltage calibrator
« Reply #74 on: October 29, 2013, 01:02:18 am »
The AN86 approach looks interesting. An similar but slightly different approach would be to PWM between 2 dacs. So same dual LTC1599 setup. You set one DAC to word N, and the other to word N+1. Then you use an analog switch to switch between the two DAC outputs, followed by an integrator/LPF. The duty cycle then determines where in the N...N+1 range you end up. And same as in AN86 you use an ADC for feedback.

Since the two input voltages for the analog switch are very close this will (hopefully) minimize the amount of switching noise.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf